[pymvpa] PyMVPA 0.5: feature wishlist

[Michael Hanke]

On Fri, Oct 02, 2009 at 12:32:52PM +0100, Johan Carlin wrote:
> I just remembered another feature I've missed in PyMVPA - this one
> might be a tad easier to implement than parallel processing. :)
> 
> There are a lot of reports that averaging your examples together
> improves classification accuracy (e.g. Mitchell et al., 2004, Machine
> Learning). As near as I can tell PyMVPA doesn't have built-in support
> for this at the moment.

Out approach is in mvpa/misc/support.py and is called

  transformWithBoxcar(data, startpoints, boxlength, offset=0, fx=N.mean)


should work. However, it is not really elegant. Such functionality will
move into proper Mappers for 0.5.


Thanks,

Michael



-- 
GPG key:  1024D/3144BE0F Michael Hanke
http://mih.voxindeserto.de